Aaron N. Tubbs bio photo

Aaron N. Tubbs

Dragon chaser.

Twitter Facebook Google+ LinkedIn Github

One of the things that I do is write a lot of notes. Things to do, things to buy, things to think about, things to try. Thoughts I have, stuff to blog about, stuff to write about, stuff to remember. Stuff to forget, random drops of information, reading recommendations, quotes for later. You get the idea. For some things, there is a best of breed place to put all of the information. For example, in terms of storing, classifying, and recalling bookmarks, del.icio.us works perfectly for me. The “folksonomies” concept that’s been floating around the blogosphere (still hurts to type) seems well suited to this sort of thing — classifying atomic chunks of same-sized data and information according to topic spaces.

For other things, post-it notes and note cards work great. “Grab milk when you go home” works pretty well on a 3×5 piece of paper. But, with large lists of not necessarily related information, when I still want to collect it all in one place I’m a little stumped as to the best approach. I’m building a free-form personal database of information in effect, and I keep finding things that aren’t optimal. For a while, I’ve been using hnb, a hierarchical ncurses outliner. In it I store shopping lists, gift ideas, and all sorts of other data that’s easy to structure. I did most of the wedding and vacation planning that I was responsible for in hnb. The thing is, the information I work with isn’t always nice and hierarchical, or the enforcing of hierarchies quickly becomes more of a nuisance than an asset.

I tried various strategies to use vim as an outliner — first using fold markers, which was effective, though somewhat irritating to maintain, and then with VimOutliner, which I found more useful, but it made things somewhat arbitrary and irritating. Integrating large free-form chunks of data to transfer to a blog entry or something else was a big pain in the ass, even with the various features built in to extract text without hierarchy information. A few more macros would have made short work of my complaints, but I found I was less using vim, and more using an application inside of vim that took a lot of the power, simplicity, and intuitiveness of vim away. Three strikes, I’ve not loaded it since. Neat in theory and a nice effort, but it’s one of those anti-products. Would fit perfectly into emacs. I guess that’s an insult.

Anyhow, my next approach is to try a wiki. Why not? I tried MoinMoin:http://moinmoin.wikiwikiweb.de/ first, but I found there were some aspects I didn’t appreciate, and the error messages that were coming back were nonsensical. It worked much better in isolation than it did under various apache configurations, though I’m sure in each case (mod_python, mod_fastcgi, normal cgi), I was just missing a setting here or there that would have fixed everything. That, and I only tried to get it working for 15 minutes; there are too many wiki products out there to actually think about the decision. Further, storage was on the filesystem, which is fine, but I’m not really running a filesystem that is optimized for gobs of little files here and there, so I next tried TikiWiki, which worked like a charm out of the box, and I’m converting the various notes repositories I have here and there into it. The nice thing is that I can access it anywhere, though I’m a little concerned about how I will access the data contained on it when not jacked in. I could always run Apache and PHP and MySQL on my powerbook, and then just have a two-way database sync process (eww), but that really doesn’t seem like the way to go; it’d be an awful lot of overhead and memory footprint to run a silly website, compared to something better suited to the task. Maybe there is somebody out there that produces a personal wiki — a self-contained wiki application with the ability to export the data in XML or something like that. I’ve not really looked for it, but if it doesn’t exist, maybe it’s time to write one … a free-form personal information manager or something like that. It must exist.

Anyhow, we’ll see how this experiment goes. I’m not sold on the idea yet, but it at least is good for cluttered data, has a bit of a hierarchy, but doesn’t enforce a painful structure, and is portable provided I’m connected.